The four laboratories participating in this program will produce and use several models of human cancer to pursue at least three problems that inspire current work with mouse models -- maintenance (defining the functions required to maintain a viable tumor phenotype), fate determination (exploring the relationship between cellular components of a developmental lineage and the phenotypes of cells transformed by mutant cancer genes), and regeneration (testing for the existence and characteristics of cells responsible for sustained growth and recurrence of tumors, "cancer stem cells"). Specific aims include: (i) To identify genetic lesions required for maintenance of the essential features of malignancy, through the design and analysis of multi-step models of human cancer, affecting lung, ovarian, and pancreatic epithelium, fibroblasts, and hematopoietic cells (B cell and myeloid cell lineages); (ii) To characterize the interaction between cells in a developmental lineage and the oncogenic events that produce tumors of different phenotypes in that lineage, with emphasis on pancreatic, breast, and hematopoietic tumors; (iii) To identify cells within such tumors that are responsible for driving tumor growth and thus replenishing tumors after therapy; and (iv) To continue to refine several existing models (e.g. for lung, pancreatic, and breast carcinomas) and initiate new models (for fibrosarcoma, multiple myeloma, and promyelocytic leukemia) that facilitate the pursuit of aims (i), (ii), and (iii). We will employ traditional means for genetic alteration of the mouse germ line (transgenes and targeted mutations), as well as other methods for gene regulation, tissue- and cell type-specific effects and somatic cell gene delivery (e.g. with avian virus vectors, tetracycline-responsive regulators and inhibitory RNAs).